Explainable AI for Decision Support in Blood Bank Operations

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Blood bank operations depend on timely decisions about inventory, blood-component availability, demand, expiry, shortages, and allocation. Artificial intelligence (AI) and machine learning (ML) can support these decisions by learning patterns from operational data, but a prediction is not sufficient on its own when staff need to understand why a recommendation or alert was produced. Explainable Artificial Intelligence (XAI) provides methods for exposing the factors, patterns, or examples that contribute to a model output. This paper examines how XAI can be positioned within decision support for blood bank operations. A systematic literature review approach is proposed using the PRISMA 2020 reporting framework. The review focuses on three connected questions: which blood-bank decisions are supported by AI/ML, which models are used, and how their outputs can be explained to human users. The paper also considers the practical requirements of explanation quality, data quality, validation, human oversight, and safe use. Based on the reviewed methodological literature, a conceptual framework is proposed in which operational data are processed by an AI/ML model, passed through an explanation layer, and presented as decision support rather than an automatic decision. The study is literature-based and does not claim experimental performance results that have not been measured. Keywords: Explainable Artificial Intelligence, XAI, Blood Bank Operations, Decision Support, Machine Learning, Healthcare AI